Abstract: Disastrous consequences following an extreme event of precipitation are related not only to its total volume and peak value, but also to its time distribution pattern. Based on the daily precipitation data of in-situ measurements and global climate model projections, this paper improves the three-segment deviation correction method, and examines the extreme precipitation events (daily extreme precipitation and the preceding and succeeding precipitation) in the Jialing River basin using a time distribution model, focusing on their historical distribution characteristics and the future trends in the frequency, volume, duration, and concentrating patterns of precipitation. The results show that the new improved method generates a more significant correction effect than that of the three-segment method, and a more effective reduction of the simulation errors against measurements. Compared with the historical period, the volumes of precipitation of two major types show an increasing trend and their durations are prolonged, while their concentration ratios take a decreasing trend. In the future projection period, the frequency, volume, duration, and concentration of extreme precipitation events all present an increasing trend followed by a decreasing trend.

Key words: low-emission scenario, event-based extreme precipitation, deviation correction, time distribution pattern, Jialing River basin